Nikon D5 full size sample photos published online

Nikon D5 camera review
Engadget Taiwan (Google translation) published their Nikon D5 review with full size sample images taken with the final firmware version (direct link to flickr gallery):

Nikon D5 實拍體驗(review)

The Nikon D5 is scheduled to start shipping on March 26th (pre-order links).

This entry was posted in Nikon D5. Bookmark the permalink. Both comments and trackbacks are currently closed.
  • mas921

    oh my oh my!! am blown away by high iso’s! 64k is usable with some post proc. noise NR, heck even right out of camera for low res social media/mobile sharing perhaps.

    • mas921

      ….and that lens has some serious focus issues…

      • longzoom

        Did not see any, so would you be more specific?

      • Spy Black

        Why a 24-120 was used is beyond me.

        • koenshaku

          with engadget shooting no less.

        • BlueBomberTurbo

          Kit lens?

          • Spy Black

            On a D5? A 24-70 is the closest I can see to a D5 “kit” lens.

        • Captain Insane-O

          Cause it’s got that gold ring lol.

        • neversink

          I agree – I would do my test shots with a 24mm f/1.4prime and an 85mm f/1.4 as well as the 14-24, 70-200 and 24-70 (all f/ 2.8) All shot in RAW + JPEG – fine to show the differences.
          In the long run, I am sure the D5 is an improvement in higher ISO IQ, but by how much? I am more impressed by the AF and increased number focus points by approximately three times.

    • where are you seeing a 64k photo, maybe I am missing it, are you talking about 6,400 maybe, not 64,000 – right? 6400 is usable on D3S, ……. it all looks good to me though

      • RMJ

        At least this one is ISO 64,000 :
        https://www.flickr.com/photos/engadgetchinese/25165488891/in/album-72157664383797109/

        The next one in the album is 32,000 (and very good quality)

        • mas921

          yup thats the one, in my phone (Note5) it looked nice (without zooming in ofcourse), so perhaps for social media (facebook, instagram ..etc) coverages or even web news, this kind of hi iso is good enough right out of the camera.

        • Horshack

          Photographs like these demonstrate the limits of Ultra High ISO imagery (>12,800 ISO), even when shot with class-leading sensors. Notice there is very little surviving detail in the image – it’s been decimated by noise reduction. The few situations where you can really get away with Ultra High ISO and still have usable detail is when the subject matter takes up a larger percentage of the frame, such as a portrait or close still life.

          • akkual

            I am quite sure those don’t have much NR applied or at all. That picture is taken in dark, which is the ultimate test for sensor (although well exposed, truest test is to get slightly under exposed dark picture at high ISO). Yet it seems to have enough detail for downscaled use and lots of DR. The lack of DR is the typical problem on high ISO, so in that sense D5 seems to be doing very well. As the amount of gathered light starts to be very little at those ISOs, we will probably never get high detail in those level of ISOs.

            • Spy Black

              If that’s an in-camera JPEG it will indeed have NR.

            • RMJ

              EXIF info tells that the camera was pretty much on default settings, almost everything is at neutral setting. Noise reduction is turned off. Though, I’m sure Expeed does some processing no matter what setting you choose at high ISOs (maybe even at low ISOs).

            • Horshack

              NR is obvious when viewing the image at 50% or 100%.

            • Eric Calabros

              It’s a chroma NR, mostly. The one DPR put in their studio tool as “Jpeg” sample, has a NR usually cleans the images in a way that we don’t see grains, as we do here. So I say its not the cleanest straight out of canera image, its the most detailed.
              Maybe half stop better than D4S

            • Horshack

              There’s a complete absence of high-frequency luminance noise in the low-light ISO 32,000 image. That’s impossible without NR.

            • Eric Calabros

              Yea, but it should be softer than this. Check imaging resources D4S samples

            • Spy Black

              Definitely luma NR going on there.

            • High ISONoise Reduction – Low

              As quoted direct from the EXIF, so therefore this is a JPEG with NR turned to ‘Low’ but not OFF.

            • RMJ

              Strange. I checked the EXIF with Xnview and it didn’t show High ISO NR (or few other settings). Must be a bug.

              But yeah, you are right, it shows up on Flickr. NR is off but High ISO NR is low.

            • Horshack

              See my reply below. The absence of any high-frequency luminance noise is a tell-tale sign of significant detail smoothing (NR).

            • Brian

              I guess we each have different ideas of usable and good DR. The shots look useless to me at those high ISOs.

          • Spy Black

            If that’s an in-camera JPEG you’re correct about NR, but how you process RAW data can yield better results, even if they’re a bit noisier. I tend to leave my images on the noisier side because A) it limits the amount of detail loss and B) the noise itself creates the illusion of even more detail.

            • Horshack

              Irrespective of the NR method applied (none, light, heavy), there will not be enough signal for a usable image at this ISO with any reasonable amount of detail discernible from noise when the (distant) subject matter detail is so small relative to the frame size as it is in this scene, at least for any largish reproduction size.

            • Spy Black

              Whatever it may be will still be better than anything that preceded it. These things are only getting better.

        • Ah, …. I missed those, …. cool, thanks!

  • Finally 🙂 /subscribe

  • Stephen Corby

    The ISO 32,000 shot looks awesome. Absolutely useable after a bit of NR in post.

    • Duncan Dimanche

      yes exactly my thought !

      • neversink

        Not bad at all. You have cleaned it up with out losing too much detail. The outer yellow leg is the exception to this. It seems way oversaturated now because of the NR, but then again it may have been over expose. I wonder if there was any detail in that yellow stand before you used NR.

  • Ned Gerblansky

    Exceptionally clean images. The ones shot at 14,400 look better than my D4 at 6400.

    • D700s

      I agree Ned. My D4 doesn’t hold a candle to this camera. Looking forward getting it and really comparing.

    • Captain Insane-O

      Did you download the full res, or look at the 300dpi scaled down ones? cause the full res images are not clean

      • Ned Gerblansky

        I downloaded a few full size images, two were at 14,400. There’s no doubt that the 14,400 on the D5 looks better than my D4 at 6400, maybe even 5000. The D4s was supposedly good for one step better and I’m guessing the D5 is good for a tleast another stop over the D4s.

        • Captain Insane-O

          Dxo claims that the d4s has acceptable iso up to 3050 and the d4 2950. That’s negligible.

          Look, all this one stop non sense is ridiculous. The d700 scored 2300, which isn’t even a half stop away from the d4s (a FAR newer camera)

          The d5 image at 14,400 has noise reduction. Full res it’s still grainy. Look at it at 300dpi and it looks great, but so does every other current Nikon ff.

          I added noise reduction in Lightroom at 1:1 to see at what level both my d750 and the d5 look similar. The d5 took 40 luminance and d750 took 35.

          Obviously this isn’t scientific, but me thinks this is indicative of the historical performance that shows Nikon first ff D3 performing within a half stop of their best current iso performer the df (2300 vs 3300 on dxo).

          I regularly use iso 6400 with noise reduction in my d750, but it loses a lot of detail. Everyone needs to decide at what detail level they deem acceptable, for post processing can handle the noise well.

          #everyCAMERAisHYPED

          • Eric Calabros

            No its not negligible. Especially at ISO 25600.

            • Captain Insane-O

              Dxo shows the graphs virtually on top of each other from 600 to half a stop over 25k. So yes, negligible noise performance difference. Certainly no where near a stop.

          • D700s

            I take it you won’t be buying the D5.

            • Captain Insane-O

              I would rather purchase an adventure bike. Haha.

            • D700s

              That’s nice.

            • Brian

              Take a look at Jared Polin’s Nikon D5 real world review (in YouTube) where he gives a download link to the RAW files. Some of the shots are a little underexposed and exhibit massive noise at even ISO 5000!

          • Duncan Dimanche

            yes I do feel the same with the recent cameras…. I think that we’ve hit a stop when it comes to high ISO performance so unless we get some new technology sensor those ISO performances will have to be done in labs to see the difference.. but in real life I don’t think that it will matter a whole lot

            • Captain Insane-O

              Yeah, problem is the same as developing countries versus developed. Already developed technology progresses much slower than newly developed tech. Look at intel, they just said there is likely no performance gains in their next architecture just battery gains

            • Brian

              I agree entirely.

          • Knut-Sverre Horn

            Well, DxO also claims that D3s has acceptable ISO up to 3253 – that is, better than D4s. I own both cameras, and I no longer trust DxO.

            • Captain Insane-O

              It’s a computer analysis, not a subjective observation. There is more to iq than the percentage of noise a file has

            • neversink

              Use your own eyes. I don’t trust DxO for a number of reasons. And human error also occurs in testing. But they do get one thing right. The image quality on a D4 is so much better than that on a Nikon 1. Thanks for nothing DxO…..

          • Originaru

            I agree, people are so hyped that want this to be the best camera HIGH ISO, im not impressed at all.

          • KnightPhoto

            Let’s not bring the D700 into this. DxO might rate it ½ stop worse but it can’t play at the level of a D4 at ISO 4000 and above. No regrets leaving the D700 behind 😉

            • Captain Insane-O

              There are other factors besides noise too. Color reproduction and dynamic range at high matter and are difficult to judge without raw format.

              But I get your point. Dxo shows a significant plummet of efficiency after 2400 iso on the d700. Current nikon ff dslr have much more linear performance, so it is hard to imagine a huge difference within native ranges.

            • neversink

              I agree, and I got rid of my D3s for your ISO reference and other reasons. However, I did keep my D700 (my wife uses it and loves it, and has become quite a photographer, actually getting some gigs on her own) and still using the D700. She is now borrowing my prime lenses including my 500 f/4 — and soon she’ll probably “borrow” the D800 and D4. Is there any better reason for me to “upgrade” to the D5 and the 600 f/4 fluoride. Now, I won’t be able to buy my Land Rover Defender. And she’s already “borrowed” my diesel Land Cruiser. Oh Well!!!!

          • osynlig fog

            I have close to zero trust in DxO.

          • Brian

            At last we have someone living in the real world! Everyone gets so excited and over-exaggerate cameras’ capabilities. I am really looking forward to the day a camera can truly capture images quality at ISO 25600 of what we get today at ISO 6400.

      • Bob Thane

        I agree – look at the ISO 3200 shot of fish skeletons in little jars – it looks like ISO 8000 on a D750. I’m hoping that’s just bad processing, I really can’t imagine the latest and greatest being a step backwards like that.

        • akkual

          There’s most likely very little processing on those files (NR “off” (it’s never completely off)). However, there seems to be variation between the shots, which is weird. There is excellent ISO25,6k (way better I have seen from any Nikon so far.. Sony A7s level stuff) then there is very wierdly processed iso25,6k as well (it can’t be no NR, as it’s like made of worms at 1:1). I assume that there still is some firmware issues and the processing is not yet final.

          • Bob Thane

            Yeah, I noticed the 14.4k shot of the tree was quite good too – as good as the ISO 3200 shot I mentioned. Who knows what’s going on, but I have hope that it’ll be impressive when we see raws.

          • Captain Insane-O

            It says nr off, but down the list it says high iso nr on low

            • Brian

              Yes, they are 2 different settings, or at least on Canon cameras they are, so I assume Nikon is the same.

  • vousplaisentezouquoi

    It looks really nice!

    • RMJ

      They are full size images : 5568 x 3712

      You need to click the down arrow (bottom right of the image) and select original resolution.

      • Brian

        These are 100% crops of images taken by Jared Polin. The jpegs were processed in LR from his RAW files – I left the images as they were… exposure-corrected by Jared but all noise reduction turned off. These high end cameras seem hardly better than consumer cameras in terms of noise levels, but better colour and contrast at those ISOs.

  • Nick

    So where are the folks that were denying my observations on high ISO now? :))

    • HF

      Without comparing to the D4s with the same lens at the exact spot as a reference point we can claim/observe anything. We don’t know about baked in processing, too. You really believe a 2 stops improvement as many others do, or what is your guess?

      • Nick

        I shoot D4’s at 12800 all the time, I know what the noise looks like in the shadows, I’ve been looking at it nearly every day for 4 years. I honestly think the D5 will be a 2 stop improvement. I saw the D3s to D4 difference of 12800 as about a 2/3 stop improvement at best.

        • Captain Insane-O

          Funny cause the images I downloaded have nearly identical levels of noise to my d750 under 12,800. They even take the same amount of luminance to have the same amount of noise. I’ll be surprised at even a 1/2 stop improvement. Especially considering the slow gains in semiconductor technology now a days.

          Btw, the images I saw has high iso noise reduction on per flickr exif data.

          With the gain in resolution counter acting some noise gains, I’m not surprised with the 1:1 results we can now see.

          • Eric Calabros

            Your D750 is good, but not D4S good. and I guess D5 is half stop better than D4S. Go figure.

            • HF

              You are comparing at pixel level, not downsized to the same resolution. Doesn’t make any sense!

            • Eric Calabros

              It’s the pixel level noise that matters for PJs. Otherwise even 5DS is also a good high ISO performer.

            • HF

              For me the normalised output is more important. Comparing different output sizes doesn’t make sense for me. Larger pixels always get more light and have a higher pixel level S/N ratio. 5Ds is good, too, but comparisons show that from ISO 6400 on even downsized images suffer.

            • Captain Insane-O

              The size of the image is what matters. Per dxo, the 5dsr is a third stop lower than my d750 when scaled to the same size image. While the d810 is marginally behind the d750 when scaled down.

              I print images that are of a certain size, so the 5dsr will have higher dpi and closer noise performance

            • Captain Insane-O

              Half stop better from where? Certainly not under 12,800. I highly doubt it will perform the same at 3200 than any ff current nikon at 2400. All these claims must be based off jpeg processing.

              I’m not claiming my d750 is better than the d5 d4 or df, but it’s so negligible that adjustment needs to be made at 1:1. Then when you scale the images down to the same size it takes a computer to decide the objective winner.

              History with dxo shows that the ff flagship have stayed at the same Snr levels but gained I resolving power. I would guess that nikon is increasing the iq enough to not harm their noise performance. If it were still a 16mp sensor, not gaining 25% more pixels, then I’d say a half stop may be probable. But you simply can’t gain resolution and pixel efficiency on a decade old refined technology in this day and age. Gaining 50% better performance is an outrageous claim for the same tech.

            • neversink

              What you say is just opinion and conjecture. We haven’t seen RAW files yet. Wait until you get your hands on one. I suspect the D5 will blow away my D4 – especially when it comes to AF abilities but also when it comes to noise reduction at high ISOs. I don’t yet know how much better the noise reduction will be. Even if it isn’t as much as people are hoping for, I’ll be happy with improvements in AF and other areas. All technology has its limitations. These limitations keep getting pushed further and further out as more research and development occurs and put into our cameras.

          • Stephen Corby

            I use a D750 as well and 32000 ISO shot looks like 12800 on the D750. I think it’ll be a serious improvement 1-1.5 stops.

            • Captain Insane-O

              Funny cause the 14,400 is worse than my d750 at 12,800.

              I simply used luminance on lightroom till they looked the same, and the d5 needed 5 more on the slider.

              Btw, high iso nr is on for these d5 shots. Regular nr isn’t on, but the high iso one makes a difference.

          • Nick

            Have you shot a D4 and looked at the images in a real world situation? Not just studio shots on a website? If you did you’d know exactly what they look like and from looking at the images on 5 different D5 samples with all the NR turned off and these images with NR applied I can see a big improvement. I also have shot my 750 that high and it’s good but DEFINITELY not as nice as the D4.

            • HF

              Sorry that I disagree. Downsized to the same resolution, I couldn’t see a difference between both cameras in terms of noise. DR is slightly better on the D4. And this is exactly in line with measurements by DXO, for example.

            • Nick

              Downsizing for comparison is ridiculous. Look at the noise at 100% that’s real world and what the camera can do unmodified. DXO has nothing to do with the real world, they are for cat shooters and doctors not for professionals making money with their tools.

            • HF

              No, it is not. When doing comparisons you should reduce as many variables as possible, otherwise how will you explain which effect causes which observed differences? Common practice in science. What sense is there in comparing an A7s to a future 100MP camera at pixel level? When doing a large wedding book for our clients or when they want to print we have a fixed output format, for example. I use my cameras such that the performance gives the best possible result after downsizing for this format. I would be very interested in participating in your superior knowledge in this matter, however.

            • Horshack

              This would only be true if you always print your images at their largest native size. Are PJ’s and sports shooters suddenly going to start printing all their images larger just because they moved up from a 16MP to a 24MP body? The reality is they’ll be printing to the same output size, which means the 24MP images will be downsampled, either explicitly by the photographer in post or implicitly by whatever software in the tool chain formats the output for print/web rendering (including the printer itself).

            • Eric Calabros

              Lower pixel level noise means more freedom for crop, which PJs do a lot.

            • Horshack

              The same thing can be accomplished with more pixels, even if the per-pixel noise improvement doesn’t scale 1:1 with the increased pixel count. Or with the same number of pixels but with improved per-pixel noise levels.

            • Captain Insane-O

              That’s not real world unless you print monstrous pictures or have an 8k monitor.

            • Captain Insane-O

              High iso nr is indicated on per exit data. It is below normal nr. High iso nr on my d750 makes a big difference, so I’m sure that is what people are seeing when claiming 2 stops gain.

          • PhilK

            Re: “slow gains in semiconductor technology”:

            Sensor improvements, by-and-large, are not primarily driven by Moore’s Law kinds of things at this point the way that various other chips that rely on incessantly increasing miniaturization are. Image sensors do not require that kind of miniaturization for the pixel array itself, unlike cutting-edge processing chips like the latest RAM and CPU chips.

            Seems to me that the majority of advances that have driven sensor tech forward in noticeable steps have been various architectural improvements such as: A) the transition from CCD to CMOS sensors which was also correlated with the shift from passive sensors and off-chip signal readout to sensors with on-chip signal amplifiers and ADCs, B) physical improvements in the microlenses and pixel elements themselves, making them closer-together and optimizing the optical characteristics overall to take greater advantage of the available photons, C) backside illumination, which further improved light-gathering efficiency by optimizing the physical arrangement of the sensor components, D) various architectural improvements to the amplifiers and ADCs, etc etc.

            • Captain Insane-O

              I think you misunderstand my intent.

              Silicon tech that’s highly refined isn’t going to gain much over time compared to emerging tech.

              I understand they’re much different. Just saying it’s slowing down. Not making magical leaps

  • D700s

    ok, I downloaded two of the images, one at 14400 and one at 25600. Both full res and ran through PS. First of all, neither had unruly noise. Yes, there was some but nothing Bridge couldn’t handle. I then bypassed bridge NR/Luminance and went to PS and just down sampled each. The little noise there is basically disappeared. Gotta say it looks pretty good. If they do have some magic sauce in camera for NR I’m OK with it. It’s working.

  • Alex Svitanak

    High iso (6400 and upper) have no common with photography. Hey, people, if you can’t shoot with 6400 it is a problem of your head (hands), not a camera. And iso 64000 will not help you. IMHO:)

    • Bob Thane

      You could say that of any arbitrary value. “If you can’t shoot with ISO 25 on a 8×10 camera, the problem is you” is equally valid. Which is to say, it’s valid from a certain point of view but also arguable. Sure, the problem is you since you should be able to bring in strobes and light it perfectly at any ISO value. But at the same time, that just isn’t realistic for a great many situations, so while true in theory it’s absurd in practice.

    • D700s

      Your comment really makes no sense. I don’t understand what you mean. We aren’t discussing creativity, composing, or hand hold techniques. We understand ISO won’t make a bad photographer good but might assist a good photographer. And why wouldn’t you want the ability to shoot cleaner images at higher ISO. It’s like the radio in your car. If you don’t like music, don’t use it.

    • akkual

      Some people need fast shutter speed to stop the motion.

    • fanboy fagz

      I agree partly. it depends on the scene though. if I could Id flash the scene and balance a bit with 5-8000 iso. I wouldnt shoot anything over that unless I had no choice, but for weddings, there is no need to shoot anything over 12800. even on the VERY dark dance floor, I use 6400 iso and flashes off camera aimed into the dance floor and it mixes well.

      I dont like the high iso shots shown here at all. very flat, crappy detail with a lot of grain. the photogs images arent that great which makes the impression look worse.

      who the hell uses 5.6 and shoots at such high iso. if you have a 2.8 lens then shoot it at that and get the lowest iso you can.

      • neversink

        These are called sample photos. Perhaps the person wanted to exhibit what the D5 could do at a higher ISO with a little more depth of field to bring out a bit more detail.

        • fanboy fagz

          no.

          • neversink

            yes

            • fanboy fagz

              wow what a comeback. you are definitely creative.

    • Stephen Corby

      Anyone who uses more than ISO 6400 must be wrong? If you think ISO 6400 is good enough for sports in covered or indoor arenas, or evening games, you just don’t know what you’re talking about. It might be easier for you to throw stones at people for using features you don’t need, but why don’t you try actually learning what people use them for before making such incorrect assumptions. These type of ill informed opinions are drive me crazy. Try a looking at high ISO shots sometime and see why they are taken that way.

      • Nick

        Right on!

      • PhilK

        I remember the days where if we pushed film to 1600 or 3200, we not only got golf-ball sized-grain and nearly non-existent shadow-detail, we were worried we’d upset the space-time continuum or something.. 😀

    • mas921

      sports.

    • What about wildlife or sports shooters shooting in terrible light? It’s their job to get the shot but if it looks like mush then there’s a problem, high iso has its place especially with longer focal lengths and if you live in the UK!

      • neversink

        I shoot wildlife often and I always try to shoot at the lowest ISO possible. I was taking photos just after sunset of an elephant and her baby – shooting at 1/50 second to keep my ISO down as much as possible. Of course, any movement from the camera or the ele would make for a lousy shot…..

    • Albert

      You don’t know what you are talking about. Let’s suppose you shoot ISO 3200 maximum, and a shutterspeed of 1/125s to freeze action. Well, in sports you are going to need at least 4x – 6x that much, meaning an exposure of ISO12800-16000. Etc.

  • The saga continues… So sad that Nikon takes this route again. The entire (nikon ) photographic world waiting to see samples from their top of the line camera and again, another set of underwhelming, improperly focused/exposed pictures taken in a parking lot by a dilettante. Why in the world don’t you guys pick a few of the world top photographers and get some images form them, like you did with the D4 campaign? I’m seriously considering cancelling my pre order for the D5. I have not seen anything that my D4s can’t do. I’m sure in the right hands is a very capable camera , but man…Major faux pas in my honest opinion.

    • akkual

      Did you even look at the pictures? Those are Sony A7s level in DR at very high ISO and when NR will be applied (those clearly are not heavily NR’d). And that is 20mpix, not 12mpix like A7s.

    • PhilK

      Nikon may have authorized this person to publish these pics but these are not, as far as I know, anything approaching “official” Nikon examples. Surely other people will also be publishing full-res D5 images as well, very soon.

  • D700s

    Are these photos from one of the sample cameras? I’m guessing they are. Will the final firmware in the production models have anymore improvement? While the D4S folks may not see a major improvement as a D4 owner I’m impressed with a few things in the D5.

    • usa

      From Peter: “sample images taken with the final firmware version”

      • D700s

        Slapping myself for not reading. Thanks.

  • School_Photographer

    For those of you who’d like to see some sample images taken with the new Nikon D5, Nikon ambassador Dave Black has been posting some images on his instagram account that are taken with the D5 and the new SB5000. His instagram handle is daveblackphoto

    • Eric Calabros

      Thanks, because Instagram resolution is way more than enough to evaluate a high end sensor

      • TylerChappell

        ROFL, exactly, right?

  • Eric Calabros

    That moment when you see the image noise at ISO 14400 is not much worse than ISO 4000

    • akkual

      And talk about the DR on that ISO14400 file. Huge!

      • HF

        How did you measure it?

        • Eric Calabros

          Just HDRized it in android Snapseed. Looks immeasurably good 🙂

  • Who shot this crap? My dog would do this well by accident. Can’t Nikon hand one of these cameras to the night-shift janitorial staff and get some images of someone emptying a bucket or smoking behind the trash cans in back of the building? Oh, and turn the AF on for them, too.

    • Eric Calabros

      Yea its crap, but now we know how this camera behaves in crap-exposed shots.

  • While a lot of people discuss the noise and noise reduction on the D5, how I wonder how many people does actually care?

    Yes of course, it would be great if the camera can really pull out a ISO 100 look from a ISO 64,000 shot. But whenever I am using my D810 from ISO 4000 upto 6400, I just don’t mind if it has a bit of noise as long as the subject looks sharp and the image renders to be usable for my needs, that’s it.

    Most of the photographers that will use this camera would probably working for newspapers, online news portals, and magazines where the largest print out size they’ll get is around an A3 size, where noise won’t be much of a deal. To be real, I never saw a reader complain that the image is “noisy”.

    I work for a publication, so I know what I am talking about. I am just tired of all the whining that humans do whenever there’s new tech.

    • Eric Calabros

      Well with the resolution newspapers or news websites need, D300 sensor performance is still good enough. So what to do? Stop making better sensors?

      • I didn’t say that Nikon should stop innovating and making new sensors. MY point was people never appreciate today’s offered technology and just keeps on complaining.

    • Bob Thane

      So if ISO 25k is good for web/print with a D4, if the D5 does a stop better you could get ISO 50k and have an equally usable image – handy in extreme cases.

    • mas921

      shutter speed! remember this is designed as a sports camera, super high iso is required for low light telephoto shots 😉

    • neversink

      I care, of course, but I will accept the camera’s limitations, whatever they are. We had to accept the limitations of film back in the day. I shot Tri-X at ISO 200 and developed 1:1 in D76 and then I used Dektol for prints, but I would dilute more or less depending on the negative and the contrast I wanted on a #2 or #3 paper. That was, of course, b&w. I gave my instructions to labs for every film (b&w or color) for my clients work. Today, we can get decent prints out fo an ISO 6400 on a D4. Not bad.
      I am buying the D5 in about six months, just in case the first few batches have production / manufacturing issues. I will be happy with the increased AF capabilities even if there were no improvements in noise and IQ at higher ISO.
      People today seem to want to take photos in near complete darkness at midnight, and have it look as if it is noon. Or they want to take a scene in a dimly lit bar and make it look like it is well illuminated.
      I prefer my dimly lit bar to look (and smell) like a dimly lit bar.

    • PhilK

      I think you have to take into consideration that there is a very specific sort of person that spends a lot of time on a blog like this reading and posting. I am inclined to think that they do not really represent the typical customer for the products.

      That’s not to say that their opinions and observations have no value, just that the majority of customers are probably going to spend less time agonizing over this technical minutiae and hair-splitting, and more time shooting pictures. 😉

  • *High ISONoise Reduction – Low

    As quoted direct from the EXIF, so therefore this is a JPEG with NR turned to ‘Low’ but not OFF.

    • KnightPhoto

      Thanks for the link to the ISO 64000, I didn’t notice that shot in the main FlickR album. It’s OK I guess, like to see more skilled use as others here have pointed out.

      • I’m going to guess these were taken by a rep or someone close and they didn’t have enough time to ‘escape’ and experiment with the camera hence the poor artistic merit in the photos.

        It would have also been nice to see some D500 shots too

  • AYWY

    Correction to admin: This is by Engadget Taiwan, or Republic of China, if you prefer. Not the usual People’s Republic of China that most people associate with when they see “China”.

    Anyway, the author concludes that it is a very capable tool for people who use it to feed their families. Focus accuracy, image quality and low-light performance are extremely capable even shooting in “idiot mode”. “Can shoot without worry” is how he describes the experience. He expects skilled professionals who need this will be happy.

    • malchick743

      Actually it should be “Engadget Chinese” which consists of reports of contributors/editors from HK, TW etc. The D5 reviewer concerned here, is a Taipei-based editor.

      But anyway, the association of Chinese websites with China is simply wrong and over-exaggerated.

    • I was looking at the URL because I do not understand anything else on that page: http://chinese.engadget.com/2016/02/27/nikon-d5-preview/

  • mas921

    “Nikon claims the D5 to offer leading high ISO performance due to color filter array tweaks that allow the sensor to capture more light than before.” http://www.dpreview.com/news/4350500600/cp-2016-nikon-shows-off-new-d5-d500-and-dl-compacts?slide=2

    • HF

      I think improvements are possible, of course. But not 2 stops.

      • Eric Calabros

        Maybe it’s better to mention what “stop” we are talking about when we talk about improvement. Sometimes half stop better raw with more powerful processor leads to two stops better jpeg.

        • HF

          Then people should clearly distinguish and mention that they refer to jpg engine performance out of camera. As there are so many parameters which can be tweaked and settings which need to be considered, I prefer raw performance comparisons using dcraw to control parameters as much as possible.

          • Eric Calabros

            Scientifically yes, but how many people use dcraw? Adobe does some terrible things with Nikon raw data, but the reality is that their raw developer is the most popular one in the whole community.

        • Elvir Redzepovic

          There are NO JPEG “stops” nor “RAW” stops.
          There are stops.
          Don’t make up things just to make your point. It makes you look childish.

          • Eric Calabros

            What?

            • Angry people you know … I get your point and I am with you on that. Probably 1stop in raw and looks like 2stops on the D5-in-camera-processed-jpeg vs D4s thanks to themes algorithm. This is useful for people that have to shoot jpg.

          • CERO

            what about traffic stops? 😉

          • neversink

            “STOP – in the name of love….”

  • Shutterbug

    Holy crap those look phenomenal considering the conditions.

  • VanHoff

    D5 Iso 6400 looks like Iso 800-1000 on the D810!!!

  • Colors and (lack of) noise look great. I hope this sensor finds its way into the Df2.

  • JXVo

    All this speculation about 0.5 to 2 stops high ISO improvement based on downloaded samples is just pointless. No conclusions can be drawn until this beast is shot side by side against D4 or D4s….same lens, settings and images….

  • Federico Gallinari

    What I can see in the photos is a little improvement in high iso, absolutely not what nikon says!

    There are some images that have a lot of noise, more than what I have seen in other camera (D3s, 1dx, A7s, A7rII).

    In terms of usability I think that the limit is maybe in <25k iso, something like 14-16k (that means 12.8 with some tips using dr).

    In terms of DR I don't think that we will see miracle in high iso, and DR means that even if you have low noise in high iso, that have no sense without color (that's why a7s give better images over competition); here all images shows a lots of contrast and many black areas.

    I'm sure that even a7rII can give 25k iso with near result, because you don't have to see resolution, thats typical from pixel peeping nerds, the images have to go on output, that means web or print, and when you go to the same output…the images will have same dimension, so you have to compare in same scale mode! (who tells that this have no sense are not a photographers but a tecno-maniacs …a long distance from this profession.

    Maybe is the first time that Nikon give specs with great distance from real performance, while Canon, that in the past have done this,…now seems to be more conservative and near to the real.

    I really hope that is the only "marketing" advice, because the camera seems to be very usefull for some kind of works (low light sports, paparazzi).
    The price is very high, maybe too much, I have to think in the upgrade because price vs difference in production is not so good, I have to wait and test it.

    PS
    very bad fault don't give the auto flickr light correction like canon! in a low light camera is an incredible cons!

  • vousplaisentezouquoi

    It is funny how so much people are so happy with *nothing*…

    Yes I have to admit that it is (a lot) better than the results obtained with other cameras.

    But for 99% of commercial use it is always crap…

    (Sorry fanboys, this is the crude reality)

  • Duncan Dimanche

    It seems that the quality was set on “fine” so those photos are JPGs right ? it is indeed pretty impressing but not crazy crazy compare to the 4s

    • Robert Slowley

      Did you look at the ISO 64,000 shot? How does that compare with the 4s?

  • Eno

    Terrible photos, I wonder how only unskilled people can lay there hands to test those camera before the release. 🙁

    • Duncan Dimanche

      those photos are not meant to be arty or anything else… just photos shot at high ISO.
      We can all agree that those photos are not good but cheers up at lest you’ve got sth to look at.
      #cuphalfempty

      • Eno

        Those images are also technically terrible shot. Man can not correctly evaluate high ISO when a bad lens is used at a soft aperture, a shaky hand, bad focus etc.

        • Duncan Dimanche

          sure

  • Duncan Dimanche

    32,000 ISO feels VERY usable !

    • Federico Gallinari

      this 8000iso is not….
      https://flic.kr/p/EiyZCY

      and this 20k…absolutely no…is good for trash

      https://flic.kr/p/Dwi1x7

      anyway sometimes the light condition are not good for a specific sensor (depend on bayer and predominant color).
      I think that d4 and d4s are not so distant, maybe we will see some improvement in jpeg and just above some iso, but…no people no million iso and no 200k, here even the 64k is terrible.

      • AC

        The ISO 20k night panorama is shot through a window/glass. It’s stated in the review and it’s evident by watching the red flare in the left part of the image. As for the night skies, it’s hard to judge. Maybe the sky itself wasn’t perfectly clear? We’d need a direct comparison with another body…

        • Federico Gallinari

          is iso 8k not a 20k..! And I’m not looking at retail, but only on the iso rendering.

          In any case we have to wait because any discussion threads are meaningless, I’m sure that without a change in technology we will not see a real improvement.

          Maybe something better in Nikon RAW because they are not real RAW, there are some pre-processing on it, so we’ll see a bit ‘ahead of the D5 and 1dx2 ..but nothing important.

          I hope in a general improvement to provide a tool that can facilitate in particular work situation (few), none of the things that we really need (because I can work even with very old camera), but a convenience.

          As always new cameras are just new toys, obviously better than the previous one, but nothing that gives us the real value of the price they ask.

          All enthusiasts are usually fans (no harm in that), or people who do not want to admit their weakness in front of the normal consumption (personally I admit).

  • SkyMeow

    Already sold my D4. Waiting for D5. These pics are no good for any real comparison but it actually shows D5’s capability. Terrible lighting condition, handheld, and in the hands of non-pro walking around just simply and quickly snap things around him without any consideration for proper settings. Under these conditions, D5 still manages to produce photos this clean. My D4 is definitely no match for D5.

  • Curtis

    Don’t get carried away by the hype. Once the RAWs are out, you will be very lucky to get a 1/2 stop improvement over the D4s. But to me a 1/2 stop would be huge, and a major achievement for Nikon.

    • D700s

      After looking at the back of my D4 many times and now handling the D5 albeit looking at the back of the camera, I’ll say it’s at least a stop, if not better. I’ll be surprised if it’s not.

      • HF

        I asked that question a few times already. How to achieve that? Have you seen all the prime sensor measurements over the past, D4, D4s, A7s, A7sii, considered best high ISO cameras at DXO, for example? The individual measurement curves show NO, I repeat NO improvement in S/N ratio at higher ISOs when downsized to the same output size. Differences in pattern noise or noise characteristics being less annoying to the human eye? Probably. Tweaking the CFA isn’t able to give you one stop more light. The SLT mirror in the Sony cams, which is taking away a lot, lot more light cost the camera less than 1/3 of a stop. Is the sensor a BSI sensor? Did QE leap? The only reason I can think of is NR algorithms improving the visual output. But very likely not sensor technology.

        • D700s

          I can live with new secret sauce in the algorithm as long as there is enough detail left in the image. Canon has lived and prospered via in-camera NR forever. Maybe Nikon found a better recipe.

        • PhilK

          Rather than constantly quoting DXO as “proof”, I think you should not so quickly and haughtily dismiss someone who says they have done something many many times with a different camera and know it well, and when using the D5 for a short-time they notice something distinctly better about it. (And they are definitely not the only individuals to have made this observation so far)

          No – the LCD on the back of the camera is not the finished output – far from it- but credit needs to be given to people who have, for example, done the same sort of thing over and over and over again with a camera with a trained eye, and have seen the results right after shooting (and have learned to interpret the correlation between the results seen on the camera’s LCD right after shooting versus the final output after RAW processing etc), and really can make a reasonable judgement of quality differential based on that experience.

          Surely within the next 1-2 weeks we will see many more samples, and the “picture” will become clearer.

          • HF

            DXO is a better proof to me, than looking at the back of an low resolution lcd, where you see a small jpg image, subject to picture profile settings and with NO direct comparison to the other camera at hand (the D5 has a much improved lcd, by the way with 2.356 Million dots compared to 900k on the D4s). Scientific measurements tell you a big part of the story, not everything of course. When looking at the available measurements, it is striking in my opinion, that improvements in S/N ratio are very small over the years. Seeing that a D4, D4s, a7s, A7sii are similar in S/N ratio gives me a hint, that it is very unlikely to suddenly get a two (or even one) stop advantage, as many people here claimed so see. Improvements in NR in software or jpg engine are more likely to be at play here. That doesn’t mean I wouldn’t like to see it, I just don’t see any sensor technological breakthrough giving this large of an improvement. I don’t think the D5 will be much stronger than an A7s in S/N ratio, I guess 1/3 of a stop at most. DR has improved more substantially when looking at the data.

            • PhilK

              First of all, I don’t understand why people here keep using this bizarre rationale that supposedly the higher-resolution LCD display on the back of the D5/D500 must ‘explain’ why the images being inspected via that LCD must somehow be ‘tricking’ the viewers into believing they are better quality captures than they actually are.

              Because in my world, when I have a higher-resolution display, if anything it more accurately EXPOSES flaws in the images displayed on it, compared to a lower-resolution/quality predecessor. 😉

              And regarding the whole question of relying on DXO or any other testing regime to TELL us whether a picture looks better or not – I once again come back to both the history of the photographic industry, and of the audio industry, both of which I have long associations with.

              I remember a time when there was no such thing as an MTF measurement. I remember a time when the concept of “Bokeh” was nonexistent and not discussed at all, as a specific subject. Did that mean we were incapable of telling apart the quality of one lens or camera from another? Of course not. One has to always guard against “Flatland’ism”: the mental self-deception that the only differences that matter (or even exist) are those which we know how to measure or quantify with some kind of measuring instrument. Hardly.

              Same thing in the audio industry. I remember a time when the industry didn’t talk about things like phase distortion, group-delay, transient intermodulation distortion, etc. Because those measurements hadn’t been invented yet. And of course when digital came onto the scene, we soon learned that simply measuring the old-school things like frequency-response, wow-and-flutter, THD etc… didn’t really tell the whole story either. (We eventually invented various new metrics.) Someone at that point who had proclaimed that digital reproduction was “perfect” using those old metrics was not much different than the people who proclaimed that the original wax cylinder in 1899 produced a “perfect reproduction of voice/music”. 😉

              So while I think it’s fine to accord DXO and other testing regimes a place in the debate, I don’t think that we can even begin to pretend we “know the whole story” just by glancing at their usual list of 4-5 numbers.

            • HF

              That is not a point of objection and why I remarked it to be a big part of the story, not the whole. I don’t talk about their overall scores, too, which I think pointless, but their measurement curves available on their site and showing S/N ratio, DR, tonality etc. vs. ISO.
              Finding a way of redistributing energy between luminance and chrominance channels, avoiding patterns etc. is an other very important aspect (part of which is related to Fuji’s x-trans concept for example). I distinguish however between improvements in hardware and software. A 1-2 stop sensor improvement (this is how people often term it) when looking at S/N ratio is not a negligible amount, it has to show in the measurements. This doesn’t tell us, whether one or the other is more pleasant to the eye, but without improved hardware, it will be difficult to follow with software to build on that. Nowadays NR algorithms paired with 2000’s sensor technology will give inferior results when compared to these algorithms used with modern sensors.

            • PhilK

              Right, but even the “unscientific” “1 stop better” is perfectly useful in the sense that, if George sees an image that visually has the characteristics that he normally sees assocated with his D4 shots at 6400 ISO, but at 12,800 ISO instead, I think it’s a perfectly valid observation. And at the very least, deserves to be investigated, rather than rejected out of hand.

              It could turn out (for example as apparently some have discovered with some of the Fuji products) that the ISO rating is “fudged”, meaning what someone thought was “2 stops better” might only have been “1.5 stops better” because a particular vendor or product has a mis-stated ISO rating by 1/2 stop.

              Tho in this particular case I don’t think it’s very likely as we are comparing Nikon to Nikon. It would be suicidal – particularly on their flagship product – for Nikon to give up accuracy and product-to-product consistency for some transient marketing advantage. It would kill them in the pro market, because people who are very careful and cautious with their craft would lose faith in their tools ability to produce what they expect them to produce.

            • HF

              Im am in line here, if people were a bit more careful in expressing it. Maybe that’s my scientific background. Oftentimes it is expressed in a too general or absolute way by commenters, in my opinion. When talking about stop I think of sensor tech not software. Maybe I am too picky, here, as people might include software+hardware implicitly.

      • faocisco

        Twice the resolution on the D5 lcd vs D4/s.
        The grain(noise) will appear less on a high resolution lcd when viewed at 100%.

        • D700s

          Good point. We’ll have to wait and see if that makes a difference.

        • PhilK

          Why would it be harder to see fine details like grain (eg luminance noise) on a much higher resolution display, compared to a lower resolution one?

  • I don’t know, something funky is going on… the dot pattern in the noise is larger than what I would expect with a 21mp image from a full frame sensor. And then really drill down some of the EXIF data… on that fifth image “Feature Showcase – Ophthalmic Lenses” display for example…

    Gain Control – High gain up
    Quality – Fine
    Active D- Lighting – Extra High 2
    Contrast +12
    Brightness -124
    Saturation +4
    Hue Adjustment -124
    High ISONoise Reduction – Low

    … like… who does that? Contrast +12 and Brightness -124? Really? And then the “History Software Agent” is listed as “Adobe Photoshop Camera Raw 9.3”, so what these have already made a round trip though a piece of software that doesn’t yet support the D5? THAT won’t mess up the image quality at all.

    All of the samples EXIF lists “Crop Hi Speed – On (5584×3728 cropped to 5584×3728 at pixel 0,0)” … that shouldn’t be “On” … that was a crop mode that was first introduced on the D2X and later morphed into the various “crop modes” on current day bodies.

    Blah… I can’t put my finger on it, but something is not right.

    • Federico Gallinari

      Maybe it depends on DL, when active, or some problem reading the new metadata.
      What I can see here is that in 100iso images with DL h2 the noise is visible and not clean as in the past with sony sensor (maybe this is from Aptina..I don’t know, I never read about nikon cmos fabric for ff), I hope is something wrong because it means lower DR at low iso…
      But..if means higher dr on high iso..why not.
      Anyway some shot without DL in low iso are very high contrast and to much black zone, and some in high iso (wo DL) have a lot of grain in the sky, like the 8k iso panorama.

      • Well… if this really is the production version of the sensor and firmware, it’s starting to feel like there’s some very minor improvement in some areas, and a few minor losses in others. Which is exactly how I would describe the new 24-70mm. A small step sideways.

        • Federico Gallinari

          I hope no because in that case I will keep the d3s and take some other bodies (at very good price now), that is absolutely on the same level of the other top (1dx, d4, d4s..and so on), and I can tell without doubt because I use them.
          In terms of IQ there are very low improvement in the last year, maybe the a7rII sensor by sony it’s a step (but have no sense resolution!! put that technology on 18-20..stop).
          In the past nikon marketing write a lot about the sensor, even about nano coating on it…this time…nothing…mah..
          I hope in a good works because the last lens are near all back steps.

          • PhilK

            I did notice they did not make grandiose claims about the sensor – except the extremely high “extended” ISO range. (Which seems to be a new record for Nikon of 5 stops above the “non-extended” limit)

  • LSZH

    D4 and D4s are pretty close together and this makes perfectly sense because the sensor is almost the same! DxOMark sensor score low light ISO difference of 109 ISO… The D810 is not too far away but has double the resolution. I think the sensor technology is available for an improvement in high ISO performance of the D5! The presently available jpg pictures are pretty useless to judge the D5 low light performance!!!

  • Originaru

    I would not use over 10k. High quality i would say until 6400

  • KnightPhoto

    That full size 32,000 looks very decent in the in focus areas. Still maybe expected better in the yellow/magenta OOF area at top-left. Better example anyways…

  • KnightPhoto

    Yeah I thought we were going to get better results in the OOF areas , e.g. sky around the moon. Tree trunk and branches seem decently detailed though for ISO 14,400. Have to look up my D4 ISO 12,800 stuff…

  • KnightPhoto

    Have to agree the ISO 64,000 looked at near full monitor size at 2048 pixels wide is pretty decent. Downloaded to CNX2 and a quick pass of NR (22,5) followed by fairly heavy and fine detail USM (50,3,4) and it does look great. Have to muck around a bit and see If I could get away with less NR.

    • neversink

      Post the results of your mucking around. Thanks. Much appreciated.

    • KnightPhoto

      Sure – here’s the ISO 64000(CNX2 Noise Reduction 22,5 USM (50,3,4) and reduced to 2048 pixels wide:

  • KnightPhoto

    Mostly very conservatively exposed high ISO stuff. Myself I’d probably punch these up a notch when shooting (i.e. I would have exposed these brighter). I wonder if he was using Highlight Priority metering?

  • neversink

    It’s very noisy, this 64K ISO image, but certainly not as noisy as I expected. I’m amazed how clear the naked woman in one of the windows turned. Such detail!!!!!

    • PhilK

      O_O

  • neversink

    No one thought there could be improvements in sensor technology after the D3s / D700 sensors. The D800 blew everyone away with three times the density level in the sensor, except for the whiners. The D800 was an incredible breakthrough in sensor improvement. And we shall se more breakthroughs in sensor technology…..

    I’ll repeat what I wrote above…

    I suspect the D5 will blow away my D4 – especially when it comes to AF abilities but also when it comes to noise reduction at high ISOs. I don’t yet know how much better the noise reduction will be. Even if it isn’t as much as people are hoping for, I’ll be happy with improvements in AF and other areas. All technology has its limitations. These limitations keep getting pushed further and further out as more research and development occurs and put into our cameras.

    • HF

      There will be improvement, there always is, but by how much? Will there be a disruption? It would be nice if there was, but looking at the improvement in sensors recently I don’t think so with this camera, as development usually starts not half a year ago. In my experience people are prone to hyperbole when it comes to a new camera. Sony is one of the premiere sensor manufacturers. They just introduced a class leading BSI sensor in the A7rii. The A7s is excellent in high ISO, too. I believe the D5 to be similar in performance, maybe a bit better. But not 2 stops better as many here seem to indicate. NR baked in in raw is a possibility, but you loose detail and could do much of it in post with good software, too. Looking at the first high ISO samples from Ole Liodden additionally showed some pattern noise, in my opinion.

  • Back to top